NP Re: Skinner

[Paul Mackin]

On Mon, 2004-11-08 at 06:44, jbor wrote:
> on 8/11/04 9:31 AM, Scott Badger wrote:
> 
> >> I can't even remember if I passed Pysch. 101.
> > 
> > Had to be a test subject, for the extra credit, to pass mine.....
> 
> I dropped out of Psych 1 because the Tests and Measurements (stats) strand
> was so mind-numbing. Came back round to it again through History and
> Philosophy of Science and, in latter times, Applied Linguistics.
> 
> But I do remember being a test subject for a friend who was doing his
> Honours year in Psych. Basic behaviourist setup: darkened booth, aural or
> visual stimuli (can't recall which), mild electric shocks if you pressed the
> wrong button or pressed too slowly. One girl I knew fled the experiment
> halfway through in tears. Another fell asleep in the booth. I guess quite a
> few of our mates just treated it as a big joke and stuffed around.
> 
> What it did illustrate was that internal reliability is automatically up the
> creek when you're dealing with human subjects, and I'd say it's particularly
> so now that there are strict ethical constraints against covert observation.
> 
> > I thought
> > Rob simply meant that (for a behaviourist) there is a causal relationship
> > between stimulus and response - "variation" only so far as our understanding
> > is limited. Usng your computer analogy, so long as all the ones and zeros at
> > the bit/cellular level are known, the output/behavior could be predicted, by
> > an all-knowing behaviourist.
> 
> Thanks, that's exactly what I meant. (I was beginning to think I was writing
> in a foreign language!)
> 
> best


I answered Scott before I read Robert's reply.

I don't know why but only now something that should have been obvious
from the start has dawned on me.

One can talk about behaviorism from two different aspects that really
never meet in practice or apparently in p-list discussions.

There is the philosophical aspect. In theory every stimulus causes a
response. Therefore if we could know every single stimulus (and no two
would necessarily ever be alike) and every single response to every
possible stimulus, we would know everything about each and every subject
that there is to know. Everything would be completely determined.

The other face of behaviorism is that of an experimental science. We are
still behaviorists. But now we must put theory into practice. We devise
a stimulus and apply it to a sample of test subjects and observe the
resulting behaviors. (behaviors and nothing but behaviors) Presumably we
will see some reasonable pattern in our results. A bell curve or
something of that order. We then summarizes our data with the most
appropriate statistical method available in order to formulate some
"law" we hope can be held to be reasonably valid within prescribed
statistical parameters.

So why was Pointsman locked in that zero/one space.

Ah me.